RE: [PATCH v3 2/4] usb: aspeed-vhub: fix remote wakeup failure in iKVM use case

[Neal Liu]

> -----Original Message-----
> From: Benjamin Herrenschmidt <benh@xxxxxxxxxxxxxxxxxxx>
> Sent: Tuesday, December 14, 2021 10:36 AM
> To: Neal Liu <neal_liu@xxxxxxxxxxxxxx>; Felipe Balbi <balbi@xxxxxxxxxx>;
> Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>; Joel Stanley
> <joel@xxxxxxxxx>; Andrew Jeffery <andrew@xxxxxxxx>; Cai Huoqing
> <caihuoqing@xxxxxxxxx>; Tao Ren <rentao.bupt@xxxxxxxxx>; Julia Lawall
> <julia.lawall@xxxxxxxx>; kernel test robot <lkp@xxxxxxxxx>; Sasha Levin
> <sashal@xxxxxxxxxx>; linux-usb@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx;
> linux-arm-kernel@xxxxxxxxxxxxxxxxxxx; linux-aspeed@xxxxxxxxxxxxxxxx
> Cc: BMC-SW <BMC-SW@xxxxxxxxxxxxxx>
> Subject: Re: [PATCH v3 2/4] usb: aspeed-vhub: fix remote wakeup failure in
> iKVM use case
> 
> On Thu, 2021-12-09 at 02:37 +0000, Neal Liu wrote:
> > I'm confused. Signaling Wakeup when wakeup_en is set if it has any ep
> > activities is not exactly what you said?
> >
> > wakeup_en is set only if host allows this device have wakeup
> > capability and bus being suspended.
> >
> > Normal ep activities would not write the MMIO because wakeup_en is not
> > set.
> 
> Hrm... I didn't think wakeup_en was limited to the bus being suspended, but
> maybe I misremember, it's been a while.
> 
> Cheers,
> Ben.
> 

wakeup_en is only set in the case of host set USB_DEVICE_REMOTE_WAKEUP feature to vhub devices.
I think this behavior only occurs during host is going to suspend, and set this feature to any device which can wakeup itself before sleep.
Thanks

-Neal